Comprehensive or specialty-specific cancer care in the United States: A story of continuing underperformance.

Abstract

6577 Background: “Textbook oncologic outcome” (TOO) is a composite measure representing the “ideal” outcome for patients undergoing cancer surgery and is associated with improved survival. Using TOO as the primary outcome, we asked whether hospitals are high-performing across multiple cancer types. Methods: Patients undergoing potentially curative breast, colon, rectal, lung, or pancreatic cancer resection were identified within the National Cancer Database (2010-2016). Organ-specific TOO was defined as: adequate lymph node yield, R0 resection, non-length of stay outlier, no hospital readmission, no 90-day postoperative mortality, and receipt of guideline-concordant chemotherapy and/or radiation. Mixed-effects analyses estimated the risk-adjusted TOO rate for each hospital stratified by cancer type. Results: Among 1,094,550 cancer resections (breast = 690,442; colon = 213,918; lung = 121,771; rectum = 40,315; pancreas = 28,104), 434 hospitals performed ≥10 resections for each cancer type. Only 11 hospitals (2.5%) ranked in the top quartile for adjusted TOO rate for all five cancer types. Of the 323 hospitals (74.4%) that ranked in the top quartile for one cancer type, 149 hospitals (46.1%) also ranked in the bottom quartile for another cancer type. There was a weak correlation between hospital rankings across cancer types with the strongest correlations between colon and rectal cancer (R2= 0.148) and lung and pancreatic cancer (R2= 0.098). Conclusions: Most U.S. hospitals do not provide high-quality care across cancer types with respect to TOO. Perhaps this knowledge should be used to guide referral for oncology care.